first time delta experience, autocalbration
-
this is my very first delta duet experience..
it works fine, it homes to max, and the z probe is triggering as supposed too, and i does mesh and calbration beatifully...
but i dont get any values anywhere! i am supposed to finetune and change lenghts, offsets, radiuses, everything in the config that says autocalbration will do... but where do u get these numbers? running mesh grid or auto calbration does not bring any number to me at all to put into config.g or save with M500....
i normally use an inductive probe on my other cart machines, which is waaaay different than a physical endstop mounted under the effector, that this one uses... it has a trigger height of approx 20mm
TIA
-
-
After you calibrate, M500 writes a file called config_override.g. You can look at that via the web interface. Look for curiosity, you don't need to adjust anything.
And/or, you can type M665 (with no parameters) in the web interface (console), and see the parameters that were just calibrated. Again, for curiosity. In fact, you can type almost any configuration M code with no parameters, to see its current state.
If you choose to save with M500, you DO need to put an M501 near the end of config.g, to make the printer read that override file when it boots.
Delta / Kossel printer fanatic
-
i have autocalibrated, but when sending "M665" alone, it shows up with the "approximate" measurements i put into config.g when i made the files on the configurator... there wwill never come any real numbers after i do an autocalibration.. i know marlin did that.....
-
my autism goes haywire on this....... i have no clue what is happening....
due to my autism, pls also respect that the stuff i can find difficult is most likely the stuff you find super easy.. an vice versa..
-
I think my sd card is messed up..... i tried rebooting the board, nothing.. still couldnt find the values....
new (different) sd card, values pops up when sending M666 and M665
great! well thnx for that!
now it just measures the rod length and endstop max positions totaally different values than marlin! rod length is like, 216.5, also when measuring, but auto calibration says 238.775
-
does this look ok?
-
@johnjohn1990 change the S9 at the end of the Bed.g file to S8 rest the rod length in config.g and try again (Then it won't adjust the rod length which is often over adjusted unless you have an almost perfect build)
-
Done a S8 autocal now
so basically i should run S6, S7, S8 and S9 as individual calibrations after eachother, ?
-
no each one does something slightly different to each other in what is adjusted if you know your arm length accurately you should stick to either S6 or S8 S8 does the same as S6 but also adjusts for tower tilts should there be any IIRC
-
-
you need to reset the M665 to have the measured rod length first
-
did that, and now i get this, which is more right, with the standard settings which made perfect sized parts..
tried a 6 and 8...
that should be enough then? i just need to run a lot to get the deviation down
-
To recap:
-
You SHOULD have a PRECISE measure of your diagonal rod length "joint center to joint center" before calibration, and put that in M665 L before calibration.
-
If for some reason you do NOT have this, you can run one of the calibrations that sets rod length... and you should expect to run it over and over. It will take as many as ten or fifteen runs to settle on the 'tradeoffs' between the various parameters.
-
The better way is to set the rod length (M665 L) to the exact physical, and run the calibrations, such as 6 and 8, that set everything else, but do NOT attempt to adjust rod length. These style calibrations should only need to be run a couple of times to 'converge', meaning get a decent deviation.
-
If deviation won't drop below 0.05 (or better) even after several runs, something else is wrong. Slop, loose belts, loose hub on a motor, something.
And... as mentioned above... after you are happy with calibration, enter M500 manually. Then be sure M501 is somewhere near the end of config.g.
-
-
@johnjohn1990 said in first time delta experience, autocalbration:
did that, and now i get this, which is more right, with the standard settings which made perfect sized parts..
tried a 6 and 8...
that should be enough then? i just need to run a lot to get the deviation down
Which firmware version are you using? There is a bug in 2.03 firmware that causes slow convergence when auto calibrating a delta. It isn't present in firmware versions 2.02 or 2.04RC1.
Duet WiFi hardware designer and firmware engineer
Please do not ask me for Duet support via PM or email, use the forum
http://www.escher3d.com, https://miscsolutions.wordpress.com -
@danal said in first time delta experience, autocalbration:
To recap:
-
You SHOULD have a PRECISE measure of your diagonal rod length "joint center to joint center" before calibration, and put that in M665 L before calibration.
-
If for some reason you do NOT have this, you can run one of the calibrations that sets rod length... and you should expect to run it over and over. It will take as many as ten or fifteen runs to settle on the 'tradeoffs' between the various parameters.
-
The better way is to set the rod length (M665 L) to the exact physical, and run the calibrations, such as 6 and 8, that set everything else, but do NOT attempt to adjust rod length. These style calibrations should only need to be run a couple of times to 'converge', meaning get a decent deviation.
-
If deviation won't drop below 0.05 (or better) even after several runs, something else is wrong. Slop, loose belts, loose hub on a motor, something.
And... as mentioned above... after you are happy with calibration, enter M500 manually. Then be sure M501 is somewhere near the end of config.g.
yep i did that
and it works! i kept the arm distance what i know it is - i cant measure more precise than the visibility of a difference in a mm scale.. i do not have calipers above 150mm... otherwise i could get it more precise.. but it is somewhere between 216 and 217mm... 216.5 is standard.. radius of 102 is also more exact, to what i have seen it can do, if i move it by hand..
could it be a shitty endstop? - using that as a probe, with the lever on it.. - i am intending to get something better..@dc42 said in first time delta experience, autocalbration:
@johnjohn1990 said in first time delta experience, autocalbration:
did that, and now i get this, which is more right, with the standard settings which made perfect sized parts..
tried a 6 and 8...
that should be enough then? i just need to run a lot to get the deviation down
Which firmware version are you using? There is a bug in 2.03 firmware that causes slow convergence when auto calibrating a delta. It isn't present in firmware versions 2.02 or 2.04RC1.
in fact i am running 2.03!
what exactly does that mean? other than i need to update, asap 
-
-
i will say i am printing right now!
so, if i ever need to change my offset, i just home, jog down to where i want my new Z0, and then do a G92 Z0 ?
as said before, im used to an inductive probe, i have never had the need to use G92 in my case might seem a bit dumb, lol.. feels like that 
and thank you VERY MUCH for giving you the time to cut it out for me
youre just great -
@johnjohn1990 said in first time delta experience, autocalbration:
yep i did that and it works! i kept the arm distance what i know it is - i cant measure more precise than the visibility of a difference in a mm scale.. i do not have calipers above 150mm... otherwise i could get it more precise.. but it is somewhere between 216 and 217mm... 216.5 is standard.. radius of 102 is also more exact, to what i have seen it can do, if i move it by hand..
Pre-made arms should have a sticker on them with xx.xx level precision.
could it be a shitty endstop? - using that as a probe, with the lever on it.. - i am intending to get something better..
Do repeated calibrations converge? Meaning below 0.05?
@johnjohn1990 said in first time delta experience, autocalbration:
i will say i am printing right now!
Fantastic!! By the way, I recommend a simple (probably 8 point) calibration before every print (if your probe supports that).
